Exploration Medical Data Architecture
Big Data Big Think Forum
April 6, 2016
Erik Antonsen MD, PhD, FAAEMElement Scientist
Sandeep ShetyeChief Data Architect
2
Exploration Medical Capabilities – Data is a Challenge
• ISS and exploration: Different Medicine requires different capabilities
• What can we learn from ISS?
• A vehicle and a ground challenge
• Risk Modeling – what are we worried about?
• Needs Analysis – what do we think we need?
• Data Architecture – Medicine as a system
3 3
4
Motivation
Medical Risk: Given that medical conditions will occur during human spaceflight missions, there is a possibility of adverse health outcomes & decrements in performance during these missions and for long term health
4
The medical system supports healthy crew to enable completion of mission objectives. We are concerned with
health and prevention, not just catastrophic events.
To minimize mission medical risk through medical system design and integration into the overall mission and
vehicle design.
5
Mission ConstraintsExploration missions are different:
- Harsher environmental constraints
- Shrinking resources (mass, volume, power, etc.)
- No evacuation option
- Limited if any resupply
- Delayed or absent communications
- The only resource growing is data handling capability (Moore’s Law)
5
Exploration may require “Stay
and Fight” Medicine, not
“Retreat” Medicine.
6
Medical and Research Data Processing• Mission Associated
Summary of Health
– Detailed Medical Report
– 500 page pdf
– Takes a team 6 months to create
• Biomedical Data Reduction and Analysis
– Around 100 pages pdf
– Preparation is pre-mission through 45-60 days post flight
Some version of both of these processes will need to be
incorporated and automated in exploration missions.
7
Challenges
• No single medical database stores all information that is complete
• Leveraging extensive quantity of medical research being published on an ongoing basis
• Integrating new data sources into existing medical systems
• Existing industry systems have limitations• Limited utilization and scope
• Design for specific purpose and difficult to personalize
• Access to knowledge base and other medical data
• Machines can assist but can’t replace humans
7
8
• Risk Mitigation Strategy– Planning
• Concept of Operations Development (Ops Risk Reduction)
– Characterization of Risk• Models and Metrics – Integrated Medical Model (IMM), MONSTR prototype
• Active Data Gathering – Medical Consumables Tracker (MCT), biosensors, Flexible Ultrasound
– Active Risk Reduction• Medical Data Handling – Exploration Medical System Demonstrator (EMSD),
Medical Data Architecture
• Technology Development – Oxygen Concentrator Module, Medical Suction, IVGen…
• Training
• Medical Decision Support
• Integration of Medical with Vehicle Designers and ECLSS SMTsMed
ical S
yste
m D
evelo
pm
ent
How do we get there?
9
Provide the crew with the
best chance to accomplish
mission and get home
healthy
Medical Operations
• Nominal Operations
• Contingency Operations
• Routine
• Urgent
• EmergentMedical System?
The Goal
Comm Delay
No Evac
No Resupply
10
Risk Characterization• Identify the medical conditions that are likely to occur.
• Identify how likely and how often they might occur.
• Identify what resources are desired to address them.
Draft Data Only
Medical Optimization Network
for Space Telemedicine
Resources (MONSTR)
Medical Capability
(MONSTR)
Event Probability
(IMM)
Notional
11
The IMM Medical Conditions
SKIN
Burns secondary to Fire
Skin Abrasion
Skin Laceration
EYES
Acute Glaucoma
Eye Corneal Ulcer
Eye Infection
Retinal Detachment
Eye Abrasion
Eye Chemical Burn
Eye Penetration
EARS, NOSE, THROAT
Barotrauma (sinus block)
Nasal Congestion (SA)
Nosebleed (SA)
Acute Sinusitis
Hearing Loss
Otitis Externa
Otitis Media
Pharyngitis
DENTAL
Abscess
Caries
Exposed Pulp
Tooth Loss
Crown Loss
Filling Loss
CARDIOVASCULAR
Angina/Myocardial Infarction
Atrial Fibrillation / Atrial Flutter
Cardiogenic Shock secondary to
Myocardial Infarction
Hypertension
Sudden Cardiac Arrest
Traumatic Hypovolemic Shock
GASTROINTESTINAL
Constipation (SA)
Abdominal Injury
Acute Cholecystitis
Acute Diverticulitis
Acute Pancreatitis
Appendicitis
Diarrhea
Gastroenteritis
Hemorrhoids
Indigestion
Small Bowel Obstruction
Pulmonary
Choking/Obstructed Airway
Respiratory Infection
Toxic Exposure: Ammonia
Smoke Inhalation
Chest Injury
NEUROLOGIC
Space Motion Sickness (SA)
Head Injury
Seizures
Headache
Stroke
Paresthesia
Headache (SA)
Neurogenic Shock
VIIP (SA)
MUSKULOSKELETAL
Back Pain (SA)
Abdominal Wall Hernia
Acute Arthritis
Back Injury
Ankle Sprain/Strain
Elbow Dislocation
Elbow Sprain/Strain
Finger Dislocation
Fingernail Delamination (EVA)
Hip Sprain/Strain
Hip/Proximal Femur Fracture
Knee Sprain/Strain
Lower Extremity Stress fracture
Lumbar Spine Fracture
Shoulder Dislocation
Shoulder Sprain/Strain
Acute Compartment Syndrome
Neck Injury
Wrist Sprain/Strain
Wrist Fracture
PSYCHIATRIC
Insomnia (Space Adaptation)
Late Insomnia
Anxiety
Behavioral Emergency
Depression
GENITOURINARY
Abnormal Uterine Bleeding
Acute Prostatitis
Nephrolithiasis
Urinary Incontinence (SA)
Urinary Retention (SA)
Vaginal Yeast Infection
INFECTION
Herpes Zoster (shingles)
Influenza
Mouth Ulcer
Sepsis
Skin Infection
Urinary Tract Infection
IMMUNE
Allergic Reaction
Anaphylaxis
Skin Rash
Medication Reaction
ENVIRONMENT
Acute Radiation Syndrome
Altitude Sickness
Decompression Sickness (EVA)
Headache (CO2)47 have occurred in spaceflight
12
Diagnosis and Treatment of Medical Conditions
Functional Impairments
ISS Medical System Resources
Risks due to Extravehicular Activities
(EVAs)
Crew Member Attributes
Mission Duration and Profile
Medical Condition Incidence Data
Medical Resource Attributes
Clinical Outcomes and Mission Impact
What should be in the Exploration Medical Kit?
What is the likelihood of a medical evacuation?
What medical devices should we have on ISS?
What is the risk of Loss of Crew Life due to illness on ISS?
Without IMM
Flight Surgeon
??
12
13
Crew Training
Integrated
Medical
Model
Flight Surgeon
Crew Member Attributes
Mission Duration and Profile
Crew Composition
Type and Quantity of all Medical Events
Risk of Loss of Crew
Risk of EVAC
Medical Resources Used
Optimized Medical System within Vehicle Constraints
Mission Specific Inputs Monte Carlo Simulations Quantified Outputs Informed Analysis
IMM Relational Database
Diagnosis and Treatment of Medical Conditions
Risks due to EVAs
ISS Medical System Resources
Medical Condition Incidence Data
13,500+ data
elements
Provides a tool to help informed decision making
13
CliFFs
Quality Time Lost
14
MONSTR
• MONSTR provides a way of quantifying a ‘Needs Analysis’
• Uses the same 100 conditions as IMM• Identifies resources ‘desired’ for diagnosis and
treatment (In an ideal world)• Ranks those resources on:
– Medical criticality and generalizability to multiple conditions
– Incorporates Probability of Occurrence– Provides a ‘value’ estimate to help prioritize research
investments
15
MONSTR - Visual Networks – DRAFT DATA ONLY
All Conditions & Resources Filtered on Advanced Airway
Advanced Airway – Worst Case Advanced Airway – Best Case
16
What’s Next?
• Once you've characterized the medical risk and assessed what you need to address it, what is next?
Put those needs together in a system designed to maximize good outcomes and minimize medical errors.
MEDICAL DATA ARCHITECTURE
17
Active Risk Reduction• A comprehensive medical system to support the crew in Exploration Missions
targeting autonomy.
• In order to do this, ExMC will need integration with vehicle systems – single point solutions are not desirable for Exploration Missions.
• Medical System
– Provide for centralized medical care
– Enhance available knowledge base
– Provide for electronic training needs
– Monitor supplies for crew
– Monitor crew as needed
– Streamline communication with ground flight surgeons
– Decrease likelihood of medical errors
18
Medical System Integrates with other needs
• Behavioral Health and Performance– Dashboard program– Actiwatch– Technology for monitoring and intervention needs to communicate with this system
• Human Health and Countermeasures– Exercise countermeasures and monitoring– Laboratory and analysis capability– These tools provide early warning for medical issues and crossover for rehabilitation needs
• Space Human Factors and Habitability– Training methods and procedures– Crew interface for the medical system– Medical Hab design
• Space Radiation– Radiation monitoring– Long-Term Health Effects
• Other Vehicle Systems– Environmental sensors (CO2 levels, cabin temperature, etc.)– Vital signs monitoring via use of vehicle cameras– Medical system integrated with vehicle data, hardware and information systems
18
19
TrainingMedical Procedures
Imaging Techniques
Behavioral Intervention
EMR
AMP
ModelsIMM
MONSTeR
Digital Astronaut
Radiation
References
Pharmacologic
Toxicologic
Medical Imaging
Training Modules
Up-To-Date TM
TrackingConsumables
Food
Medications
Fluids
Medical Eq.
SensorsBiomonitors
Environmental
EVA
Exercise
Behavioral
PerformanceBehavioral Monitor
Exercise Monitor
EMSData Streaming
Medical Decision Support
Telemedicine
Autonomy
Semi-autonomy
Intelligence Augmentation
Medical Equipment
Devices3-D Printing
Medications
Imaging
Rehab
Displays /
Conferencing
Caregiver
Interface
Vehicle Data
Interface
Medical System Capture Diagram
20
Data Architecture and Medical Data Analytics
• Data Architecture
– Capture and collect from disparate data sources
– Transform and integrate all data
– Provide uniform access to data
– Build analytics datamarts
• Analytics of patient health data
– Monitor data
– Process data and infer adverse health events
– Aid crew medical officer in a broad range of diagnostic tasks
– Enhance crew safety
20
21
Data Transformation Model
22
ExMC Data ArchitectureData Sources Layer Data Storage Layer Analytical Layer User InterfaceDiscovery & Analytics
Structured
Unstructured
Streams
Health Records
Medical Records
Clinical Trials
Other ….
Medical devices
Monitoring System
Images
Logs & Notes
Exercise Machine
Bio Sensors
Env. Sensors
Other ….
Other ….
HL7 –
Adapte
rs –
Messagin
g B
us –
Data
ET
L
EHR Documents
Sensor Other
Vitals
Data Assets
Data Models
Annotate
Correlate
Classify
Integrated Data
Platform
Clinical Decision SupportSystem
Data
Serv
ice
s
Analytics Data Mart
Relational
Data Mart
non
relational
ReportsDashboard
Data MiningText Classification
Computational Statistics
Modeling & AnalyticsDiagnosticPredictive
DiscoveryOntological Search
Real Time AppsAlerts
Cognitive ComputingAdaptive, Interactive,
Contextual
AP
I, I
nfo
rmation S
erv
ices
Applications& Prototypes
User Interface &
Visualization
Metadata & Data Standards
Data Virtualization
Federated Access & Delivery Infrastructure (FADI)
Knowledge Base
Knowledge
Models
23
Medical Decision Support System (MDSS)
A knowledge system designed to use patient medical data and medical knowledge to generate case-specific assessment and
recommendations to help medical staff make medical decisions
23
Collaborative effort with Canadian Space Agency
24
Hybrid Approach for Implementation
• Knowledge based• Use of knowledge bases
• Inference engine
• Decisions based on rules
• Non-knowledge based• Machine learning
• Neural Networks (ANN/CNN) and algorithms
• Derive knowledge from patient data
• Learn from decision trees
24
25
Medical Decision Support Module
Medical & health records
Bio & Env. sensors
Images
Test results
Training
References
…..
Data Integration and Transformation
Information Interpretation
Health Assessment &
Predictions
Medical Decision Support Module
Machine learning & Patterns
Knowledge base & Inference engine
26
Medical System Development
• Initial Focus is on Use Cases defined with the Clinical Team
• Later development is driven by the Concept of Operations
• An early testbed with a subset of systems is created from use cases
• An incremental and iterative testing pathway is created to build up the system
26
27
Vehicle Exploration Medical System
Ground Based and Vehicle Data Architectures:
• Clinical Operational Needs
• Research Data Capture
• Long Term Health Information
• Crew Medical Officer
• Crew Medical Support
• Flight Surgeon/BME
• External Consults
Medical Data Architecture
Real-Time Data Processing for Crew
Mirrored Delayed Data Presentation for
situational awareness/support
28
EMR
LSAH
SMOT
PMC
LSDA
Other
T2
Other
Ad-hoc Datasets
CO2
Cardio
VIIP
Scheduled Datasets
Exploration & Discovery
Reports & Dashboards
Analytics & Visualization
Content Analytics
Machine Learning
IDENTIFY DECIDEACQUIRE
Extract, Transform & Load Data
Cleanse Data
Data Profile & Verification
Exception Handling
Workflow & Rules Management
Batch Processing
Contextual Datasets
EMR LSAH SMOT
PMC LSDA Other
Operational & Transactional Data
Historical & Trend Data
Reference & Meta Data
Audit & Exception Data
Processed Data sources
Data Granularity
ORGANIZE
Realtime Stream Processing
MASH Report
Flight Surgeon
Dashboard
Data Catalog &
Search
Security & Access Control
Data Governance & StewardshipCurrent Processes &
Protocols
Querying, Mining & Exploration
Controlled Data Modification
De-Identification
Data sharing & Consent Workflow
Exporting to multiple formats
Scheduling Batch Jobs
ANALYZE APPLICATIONS
Store datasets as source data
Portfolio Analytics
Ground-based Data Architecture
29
MedBSharepoint Site
LSAH (lab results & Nutrition Data)
Archive data (on FTP site)
MASH report
Summary Analysis(manual)
2013 2014 2015
MR032L MR031L MR078L
MR051L MR032S
MR051L
MR051L
MedB test timeline by Crew Member
Example – JFK assassination timeline
Colors represent the type of test – Immunology, Env. Monitoring, EVA, Bone & Calcium Physiology
Sample Visualizations
29Return
30
Summary
• The ExMC research plan builds on what is known about medicine and physiology in spaceflight to – Characterize the likely medical risks– Identify medical needs to address those risks– Map out a medical system to optimize crew response to those risks
• Medical Data Architecture• Medical Decision Support to Crews• Interface with Ground Data Architecture for clinical and research needs
– Engage in a testing pathway to validate and improve that system
• This necessitates integration with vehicle design concepts early in the design cycle
• It’s all about the DATA and knowledge gleaned from data that enable good decisions
• Data Architecture and Data Management approaches are keys to Mission Success
31
BACKUP
SLIDES
32
System Approach Reduces Medical Risk
• Safe and Effective Pharmacy
• Oxygen Delivery
• Medical Suction
• Ultrasound Imaging
• Laboratory Analysis
• Biosensors/EKG
Ris
k
Capability
1 2 3 4 5
FY14 FY15 FY16 FY17 FY18 FY19 FY20 FY21 FY22 FY23 FY24 FY25 FY26 FY27 FY28
Milestone SlipGroundRev PPBE17
FlightRev PPBE17
ISS Required Integ.Rev PPBE17
ISS Required PRR Integ.Rev PPBE17
ISS Not-Req. PRR Integ.Rev PPBE17
Ext-Deliv (ISS Req) PRR Rev PPBE17
Ext-Deliv (ISS Not-Req)PRR Rev PPBE17
ISSDRM
Non-ISSDRM
Page 1 of 3 2/16/16
Exploration Medical Capability
MEDICAL RISKRev-PPBE17 + FY16Q1 Updates
01
02
03
04
05
06
07
08
ISS 1YMAsteroidPhase A
CCP EM-2 AARMEM-3. .EM-4
EM-5ISS End
EM-6 (ARCM) Mars Phase AISS End
Initial Concept of OperationsIntegratedMedical System
ConOps forall DRMs
PharmacyRecommendation
SelectTechnologies
OptimizedMedical System
ConOps Mars-Orbital Draft OMB
Gap Closure Assessment
Gap Closure Assessment
Gap Closure Assessment
Gap Closure Assessment
Clinical Outcome Metrics (NSBRI)
Assisted Medical Procedures
Gap Closure Assessment
Gap Closure Assessment
Medical Data Architecture
Exploration Medical System
Biomedical Platform (NSBRI)OMB
Gap Closure Assessment
MONSTR
Integrated Medical Model (CHS; Tool)
Physician CMO
Gap Closure Assessment
We do not have a concept of operations for medicalcare exploration missions.
We do not have the capability to provide a safe andeffective pharmacy for exploration missions.
We do not know how to apply personalized medicineeffectively to reduce health risk for a selected crew.
We do not have a defined rehabilitation capability for injuredor de-conditioned crew members during exploration missions.
We do not know how to train crew for medical decisionmaking and medical skills to enable extended missionor autonomous operations.
We do not know how to define medical planning or operationalneeds for ethical issues that may arise during exploration missions.
We do not have the capability to comprehensivelyprocess medically-relevant information to supportmedical operations during exploration missions.
We do not have quantified knowledge bases andmodeling to estimate medical risk incurred onexploration missions.
Mapped to Stability Risk
Notional
Notional
Notional
Medical System
33
Spaceflight Medical Events Risk
~100 Medical Events
in IMM
~30 Specific
Human Risks
Aerobic Capacity
Dynamic Loads (overlap)
Electric Shock
HSDI
Immune (some overlap)
Micro (some overlap)
Muscle
Nutrition/Food
Neuro-vestibular
Long Term Osteoporosis
Medication Storage
Orthostatic Intolerance
Radiation (long term)
Sunlight Exposure
Team Cooperation
Abdominal Injury
Abdominal Wall Hernia
Abnormal Uterine Bleeding
Appendicitis
Acute Diverticulitis
Anaphylaxis
Eye Abrasions
Dental
Diverticulitis
Gall Bladder
Hernia
Indigestion
Med Overdose/Reaction
Pancreatitis
Sepsis
……
…..
Acute Radiation
Altitude Sickness
Cardio
Angina/ Myocardial Infarction
Atrial Fibrillation/ Flutter
Stroke
Sudden Cardiac Arrest
Back Pain
Behavioral/Depression
Bone Fracture
Carbon Dioxide (headache)
Celestial Dust Exposure
Decompression Sickness
EVA Injuries
Hearing Loss
Renal Stones
Sleep
Specific Toxic Exposure
Urinary Retention
VIIP
Top 100 In Mission Medical Events and 30 Specific Human Risks
34
Medic
al R
isk
~100 Medical Events in
IMM
~30 Specific Human
Risks
Spaceflight Medical Events Risk
Includes the ~ 100 IMM medical events which includes the 30 specific Human Spaceflight risks
This risk will integrate the likelihood and consequence of all human spaceflight risks for
each DRM
Medical Conditions for which we have not planned.
Spaceflight Medical Events Risk
35
• Selection
• Screening
• Primary Prevention
• Vehicle Design Standards
• Mission Architecture
L
C
• Secondary
Prevention
• Diagnosis
• Treatment
• Chronic
Management
• Rehabilitation
Medical
Capability
For known risks:
How do we decrease this?
Keep it from happening?
But what if it happens?
Risk Characterization
36
Challenges
• Networked Complexity• Communication
– Centers– Elements– Outside HRP – SD, EA, HSRB, OCHMO and many others
• Increment and Iterate– Modified agile system development approach needed– Flexibility in approach
• Optimization pathway– The solution today is likely not the solution in 10 years– Flexibility in architecture– Standards definition and upgrades
37
How are resources ranked?
• Summed across all conditions
𝑀𝐶 = 1
𝑖=1
100
𝐶𝑅1𝑖 + 2
𝑗=1
100
𝐶𝑅2𝑗 + (3)
𝑘=1
100
𝐶𝑅3𝑘
𝑊𝑀𝐶 = 1
𝑖=1
100
(𝐶𝑅1𝑖) 𝑝𝑖 + 2
𝑗=1
100
(𝐶𝑅2𝑗)(𝑝𝑗) + (3)
𝑘=1
100
(𝐶𝑅3𝑘)(𝑝𝑘)
Mc = Aggregate Criticality Score
WMc = Weighted Aggregate Criticality Score
CR1 = Resource Criticality Score of 1
CR2 = Resource Criticality Score of 2
CR3 = Resource Criticality Score of 3
P = Probability of Occurrence
38The Concept of Operations Drives System Design Notional
Gap Conceptual Flow
(Med08)
39
Med01 We do not have a concept of operations for medical care during exploration missions.
Med02 We do not have the capability to provide a safe and effective pharmacy for exploration missions.
Med03 We do not know how we are going to apply personalized medicine to reduce health risk for a selected crew.
Med04We do not have a defined rehabilitation capability for injured or de-conditioned crew members during exploration missions.
Med05We do not know how to train crew for medical decision making or to perform diagnostic and therapeutic medical procedures to enable extended mission or autonomous operations.
Med06We do not know how to define medical planning or operational needs for ethical issues that may arise during exploration missions.
Med07We do not have the capability to comprehensively process medically-relevant information to support medical operations during exploration missions.
Med08 We do not have quantified knowledge bases and modeling to estimate medical risk incurred on exploration missions.
Med09We do not have the capability to predict estimated medical risk posture during exploration missions based on current crew health and resources.
Med10 We do not have the capability to provide computed medical decision support during exploration missions.
Med11 We do not have the capability to minimize medical system resource utilization during exploration missions.
Med12 We do not have the capability to mitigate select medical conditions
Med13We do not have the capability to implement medical resources that enhance operational innovation for medical needs
Medical RiskNew Gaps
40Relationships between research divisions in the Exploration Medical Capability Element.
Pedigree:
DoDAF
Incremental and Iterative Approach
How to decompose the work
41
Background
• Exploration Medicine is unique:– NO regular resupply of materials
– NO real-time communications
– NO potential for evacuation if serious medical concerns arise.
• Medical care includes:– Screening
– Prevention
– Diagnostic capability
– Treatment capability
– Follow up care and Rehabilitation
– Prognosis
• Characterize the likely medical risks
• Identify medical needs to address those risks
• Create a medical system to optimize crew response to those risks
• Engage in a testing pathway to validate and improve that system
• Work with vehicle engineers and flight surgeons to ensure useful implementation of that system
Exploration may require “Stay and Fight” Medicine, not “Retreat” Medicine.
42
ExMC Responsibilities
Risk Title: Risk of Adverse Health Outcomes & Decrements in Performance due to Inflight Medical Conditions
Description: Given that medical conditions will occur during human spaceflight missions, there is a possibility of adverse health outcomes and decrements in performance during these missions and for long term health.
Other ExMC Risks:1. Risk of bone fracture due to spaceflight induced changes in bone.
2. Risk of ineffective or toxic medications due to long term storage.
3. Risk of renal stone formation.
Top Related